The traveling-wave tube (TWT) has been developed to where it is a very useful device operating over very great bandwidth and at high efficiency. The TWT employs an electron beam in association with an RF wave on a slow-wave electrical structure so that a substantial interaction occurs between the RF wave and the electron beam resulting in the RF wave being amplified. when the electron beam velocity is adjusted for maximum interaction, it is found that large amplification ratios can be achieved at reasonable beam values in reasonably small structures. Since extremely broad band slow-wave structures can be achieved and, since this is the only bandwidth limiting element in the tube, extremely broad-band amplifiers have become available.
Unfortunately, for such a TWT to operate efficiently, it must be operated at high RF levels. A high level signal entering the slow-wave RF structure will cause bunching of the electrons in the beam near the output end of the TWT because, as energy is transferred from the electron beam to the RF signal on the slow-wave structure, the electrons slow down in clusters in synchronism with the RF wave motion. If the RF input level is further increased, by the time the electron beam reaches the output end of the slow-wave structure it is completely bunched. For this condition further increase in the RF input signal will not result in a corresponding increase in output power. This condition is called "saturation".
For nearly complete bunching of the electron beam, it is clear that the bunching, as a function of applied energy, will be very nonlinear and the process will thus favor the generation of harmonics. For example, in a two-octave TWT, the lower frequencies are represented by at least four harmonics that fall inside the passband (the fundamental through the fourth harmonic thereof). At saturation the TWT will be capable of a particular power output value which is set by the beam power input multiplied by the tube efficiency. If the tube itself generates harmonic energy at its output, the output at the fundamental will be degraded thereby. For example, assume that a particular tube produces an output of 100 watts and that, due to its nonlinearity, the following harmonics are generated:
Harmonic % of Fundamental Power (Watts) ______________________________________ Fundamental 100 34.5 2nd 81 27.8 3rd 81 27.8 4th 28 9.9 Total 100.0 ______________________________________
It can be seen that if the tube produced no harmonics it would develop 100 watts of fundamental output. As it is, it only produces 34.5 watts or only slightly over 1/3 of its capability, a loss of almost 5dB. If the 2nd and 3rd harmonic power being developed in the tube could be eliminated, the output could be raised from 34.5 watts to 90 watts, an increase of over 4dB. This reduces the harmonic power being developed to less than 1dB.
The technique of harmonic conditioning has been developed to offset the above described effect of harmonic generation. In its simple form this consists of feeding a distorted signal (a signal which has harmonics) to the TWT. The distortion must be in the form of a harmonically related signal that is out of phase with the internally generated harmonic signal when it arrives at the TWT output. Thus the induced harmonic is used to offset or cancel the harmonic being internally generated. One method for doing this is to apply the TWT input to a frequency multiplier, the ouput of which is passed through a phase shifter and fed into the TWT. The phase shifter is set to produce the harmonic input signal phase that will cancel the locally generated harmonic near the output end of the TWT. An alternative approach is to drive a power amplifier TWT with a TWT driver. The tubes are designed so that both are normally operated near saturation. Thus the driver TWT will generate harmonics near the output end of its slow-wave structure. Then a harmonic phase shifter is interposed between the two TWT's. This phase shifter has a phase shift-versus-frequency characteristic that is linear and of a magnitude that causes harmonic phase inversion. Thus the harmonics generated in the first TWT are phase inverted so as to cancel the harmonics generated near the output end of the slow-wave structure of the second TWT.
The above approach to harmonic cancellation provides a substantial improvement in TWT performance but does not provide anything like the performance that should be available because it only conditions the phase of the harmonic signal being injected.